Reading, an essential skill for successful function in today’s society, is a complex psychological process involving vision, memory, and language comprehension [1, 2]. Variability in fixation durations during reading reflects the ease of text comprehension [3, 4, 5], and increased word frequency results in reduced fixation times [6, 7, 8]. Critically, readers not only process the fixated foveal word but also preprocess the parafoveal word to its right, thereby facilitating subsequent foveal processing. Typically, text is presented binocularly, and the oculomotor control system precisely coordinates the two frontally positioned eyes online [7, 9, 10, 11, 12, 13]. Binocular, compared to monocular, visual processing typically leads to superior performance [10, 13, 14, 15], termed the “binocular advantage”; few studies have investigated the binocular advantage in reading [16, 17, 18]. We used saccade-contingent display change methodology  to demonstrate the benefit of binocular relative to monocular text presentation for both parafoveal and foveal lexical processing during reading. Our results demonstrate that denial of a unified visual signal derived from binocular inputs provides a cost to the efficiency of reading, particularly in relation to high-frequency words. Our findings fit neatly with current computational models of eye movement control during reading, wherein successful word identification is a primary determinant of saccade initiation.